Design And Implementation Of Face Recognition Technology Based On Video Data In Company Attendance System

Company attendance is one of the necessary methods of company management,which can effectively supervise the on-time and off-time of employees,and ensure the smooth progress of work.At present,the conventional time and attendance methods are mostly humanviewed documents,fingerprint-based time attendance machines,and punch-in APPs.With the company's need for a perfect management system,the reliability,authenticity,and security of time and attendance data are increasingly valued.Because of the advantages of biometric technology,the face recognition attendance system has gradually become an alternative solution,and many large Internet companies have implemented face recognition attendance methods.However,many companies still use the traditional time and attendance method,which defeats the enthusiasm of the employees,can not guarantee the reliability of the time and attendance records,and wastes many human resources.In recent years,with the appearance of more diverse deep learning network structures and the improvement and richness of data sets,researchers have made great progress in face recognition in still images.Based on this,this article designs and implements video Face recognition company attendance system,through the installation of a camera device in the company's entrance area,using dynamic face recognition technology,the company employees' attendance completion records.The system uses the Qt framework as a development platform to complete the implementation of the administrator module,face registration module,face feature extraction module,attendance record module,and data export module.The lightweight and portable SQLite is used as the database to store information.The core algorithm of the face recognition module uses MTCNN & FaceNet to improve the recognition accuracy and speed of the algorithm.MTCNN is used as the front end of the module to complete the tasks of identifying faces,locating faces,and aligning faces.FaceNet is used as the back end To complete the task of extracting facial features.The video stream face recognition system implemented in this article is in line with the actual application requirements,and combines artificial intelligence and company management.The system uses the webcam and non-webcam as the information input terminal of the attendance system,loads face recognition algorithms,displays attendance records and other functions,and displays real-time results.After testing and verification,the system's various functions have reached the expected results,which proves that the system The effectiveness and practicability of the system.